swapctl

A simple Go tool to manage Linux Swap space. It helps you create a swap file, configure partitions, and tune kernel settings like swappiness and cache pressure.

One-Line Installation

Run the command for your system architecture:

For AMD64 (Most VPS and Servers):

curl -L https://github.com/taha2samy-3/swapctl/releases/download/v1.1.1/swapctl-linux-amd64 -o swapctl && chmod +x swapctl && sudo ./swapctl

For ARM64 (Raspberry Pi, AWS Graviton):

curl -L https://github.com/taha2samy-3/swapctl/releases/download/v1.1.1/swapctl-linux-arm64 -o swapctl && chmod +x swapctl && sudo ./swapctl

---

Features

• Detects available partitions and free space.
• Creates and enables swap files automatically.
• Updates /etc/fstab for persistence after reboot.
• Tunes vm.swappiness and vm.vfs_cache_pressure.
• Safely checks for existing swap files before overwriting.

Usage

When you run the tool:

1. Select the partition number.
2. Enter the swap size (e.g., 2G or 4G).
3. Set the swappiness level (Recommended: 10).
4. Set the cache pressure (Recommended: 50).

Requirements

• Linux OS (Ubuntu, Debian, CentOS, etc.)
• Root/Sudo privileges.

Workload Tuning Recommendations

Workload / Service	vm.swappiness	vm.vfs_cache_pressure	vm.overcommit_memory
Redis / In-Memory	1	100	1
SQL Databases	10	50	0
Web Servers	60	100	0
High I/O Storage	10	100	0

---

Understanding Kernel Parameters

1. vm.swappiness (0 - 100)

Controls how aggressively the Linux kernel moves memory pages from RAM to the swap space.

• Low Value (1-10): Instructs the kernel to avoid swapping as much as possible. It keeps application data in RAM until memory is critically low. Best for databases to prevent disk I/O latency.
• High Value (60-100): The kernel will swap idle processes out of RAM more frequently to free up space for file system caching. Good for general-purpose desktop or balanced web servers.

2. vm.vfs_cache_pressure (0 - 100+)

Controls the tendency of the kernel to reclaim the memory which is used for caching directory and inode objects (metadata about where files are on the disk).

• Low Value (e.g., 50): The kernel retains metadata in RAM longer. This consumes more memory but significantly speeds up file lookups and disk queries. Highly recommended for SQL databases.
• High Value (>100): The kernel aggressively drops metadata from RAM. This frees up memory quickly but slows down file system operations. (Default is 100).

3. vm.overcommit_memory (0, 1, or 2)

Defines how the kernel handles large memory allocation requests from applications.

• 0 (Heuristic - Default): The kernel estimates if enough memory is available. If the request is absurdly large, it denies it.
• 1 (Always Overcommit): The kernel pretends there is always enough memory and grants all requests. This is mandatory for Redis because it uses a background process (BGSAVE) that temporarily forks memory. Without this, Redis will crash with Out-Of-Memory errors.
• 2 (Strict): The kernel strictly checks RAM + Swap limits and denies any request that exceeds them. Excellent for system stability but can cause strict applications to fail.

---

Technical Justifications

• Redis / In-Memory: Swap latency destroys Redis performance, so swappiness is set to 1. overcommit_memory MUST be 1 to allow background snapshot saving without crashing.
• SQL Databases (MySQL/PostgreSQL): Lowering vfs_cache_pressure to 50 ensures fast disk metadata lookups. swappiness at 10 prevents the database's internal memory buffers from being paged to the slower disk.
• Web Servers (Nginx/Apache): The default values (60, 100, 0) provide the best balance between serving active connections and keeping the server responsive.
• High I/O Storage: A low swappiness (10) ensures core networking processes stay in RAM, preventing transfer drops, while keeping other values at standard defaults to manage large file blocks efficiently.

License

Open Source - MIT. Created by Taha Samy.